Inspiration - We think it would be awesome and useful for an application to take a potential tweet from a user and suggest them tweets!
How it works - After scraping data from recent tweets, we use the Tf-Idf method of language analysis to find the top three most important words - keywords - from the users input. We then go through Twitter's data again comparing recent tweets with each keyword to find other similar hashtags. In the end, we have 30 somewhat relevant hashtags. Of course, given a limited data set allowed by Twitter, we could not run a full-proof analysis, leading sometimes to humorous results. But we believe this service could truly be useful for small businesses trying to maximize their reaches.
Challenges we ran into - The two large programming portions - the language analysis and the hashtag generator - were fun and interesting problems, but the true challenges came when putting everything together and in the attempt to make our product user-friendly.
Accomplishments that we're proud of - The first time, when inserting a potential tweet into our website, that we got out relevant, interesting hashtags was a great experience! Seeing the different complex parts working together to produce a cool product has been really fun.
What I learned - (Charlie) a TON about the TwitterAPI and more about the ins-and-outs of Python. Also a lot about language analysis - (Alex) Javascript, Statistical Analysis, and more about Python
What's next for TweetTime - Lots of touch-ups and making the user experience better. The main engines of the program need to be tweaked, but more time will likely be spent making sure that it is simple. If we decide to invest more in the idea, we will definitely be buying data through Gnip so that we can have truly enough access to the data to make high-quality decisions. If we get to that stage, we will definitely be making more robust statistical analysis tools as well.
Log in or sign up for Devpost to join the conversation.